The 2002 Bali bombing remains the largest single act of terrorism to have directly affected Australians.1 Such atrocities pose serious challenges to the resilience of survivors, and to the health services charged with their care. Raphael and colleagues highlighted the limited available evidence to guide practice with survivors of the 2002 bombing, including those bereaved by the attack.2 These individuals may experience the interplay of trauma and grief symptoms and have complex, evolving needs over time.2,3

Since the attacks on the World Trade Center and other United States targets on 11 September 2001, there has been an increased research focus on understanding the recovery trajectories and health risk factors of terrorism survivors.4 However, virtually no studies have sought the views of survivors themselves regarding factors that support optimal coping and recovery. Such advice can uniquely inform emergency and health planners about survivor needs.5 Drawing on the experiences of Australians directly affected by terrorism, the aims of this study were (i) to identify factors that would support the coping of similarly impacted groups, and (ii) to determine whether advocated coping supports vary according to demographic, physical and mental health, incident-exposure and bereavement variables.

Of the 81 individuals contacted, 55 (68%) participated. Of these, 32 were women, and the mean respondent age was 50 years. Twenty-one respondents were present in Bali during or soon after the bombing, and 54 were bereaved due to the incident. Respondents provided a total of 114 comments. The average interval between the 2002 bombing and the interview was 7 years and 11 months. No significant differences were found between the respondent group (n = 55) and the total Bali Recovery Program population (n = 115) regarding mean age (P = 0.38) or sex (P = 0.39).

Our findings highlight several areas that can inform both the type and targeting of support initiatives, and current policy regarding access to health and other services. Receiving professional mental health services was the most frequently advocated coping support process and did not differ significantly based on demographic, health, exposure or bereavement factors. This suggests that when services are proactive and event-specific their tolerance and perceived benefits may be similar across future recipient groups.

It was notable that women were significantly more likely to advocate proactive government responses, while men generally saw personal attitudes and responses as pivotal to effective coping. These outcomes may reflect broader findings regarding sex-related stress coping, with men and women tending to employ problem-focused and emotion-focused (affiliative) coping strategies, respectively.12 Such findings may have implications for target interventions. For example, while no differences by sex were evident regarding the advocacy of counselling per se, a small number of men expressed preferences for counselling or group work focused on symptom education (what to expect) or flexibly delivered contact (eg, a periodic check-up). As men are also more reluctant to attend mental health services in general,13 future survivors may benefit from comprehensive stepped-care services that include brief intervention and education options.14

The strong advocacy of social support processes is consistent with other findings of an association between perceived support and lower mental health morbidity among survivors.15 Disaster losses and specific trauma effects (eg, withdrawal, depression) often degrade interpersonal relationships over time. The ability to actively maintain social support in such contexts may be highly protective.16 The finding that survivors with a mental health diagnosis were significantly less likely to advocate social supports may indicate that they perceive their available supports to be insufficient or that their capacity to access them is compromised due to mental health related issues. Support maintenance may therefore constitute an important early intervention. A recently developed program, Skills for Psychological Recovery, teaches disaster survivors key recovery skills (eg, managing reactions, problem solving, rebuilding social supports) and could be adapted for terrorism survivors.17

This study has several limitations. Its findings are cross-sectional, and coping support perceptions may change over time. The response rate could potentially introduce responder bias, although no significant differences by age or sex were observed between the study and total program samples. Participants had varying contact with a therapeutic program, which may have inflated their advocacy of counselling. Importantly, such advocacy reflects direct experience of this process and its perceived value in this context. While qualitative data are often context-dependent, these findings are likely to be relevant to other Australians directly affected by terrorism and the policy and practice frameworks intended to support them.

3 Prevalence of advocated coping supports, by subthemes (n = 55)

In the late 19th century, the founder of modern sociology, Émil Durkheim, showed that suicide rates differed between countries.2 He found that, in Europe and Scandinavia, the order for countries ranked from lowest to highest suicide rate was Italy, England, Norway, Sweden and France, and recent research has confirmed that this order has remained the same.3 The consistently different suicide rates of different countries suggests that culture influences these rates. The impact is presumably bidirectional — culture (in essence, the shared products of human society4) shapes our lives and, over time, our lives shape culture.

Opera is one of the many strands of aesthetic culture in human society. Monteverdi’s L’Orfeo (Orpheus), which was debuted in 1607, is regarded as the first successful opera; about 30 000 operas have been written subsequently, most never staged.5 Fortunately, the texts (libretti) of produced operas are available in print and amenable to examination. Thus, apart from having artistic appeal, operas are able to provide snapshots of librettists’ ideas about emotions and reactions to adversity at particular times in history. It has been argued that opera, like other art forms, “reflects contemporary beliefs about interpersonal relationships and society”.6

With the task of entertaining an audience, which may involve a degree of exaggeration and the expectation that good will triumph over evil, the librettist generally strives to present a story which is plausible to the audience. To this extent, opera has the potential to offer a valid perspective of the attitudes toward suicide at specific times.

The topic of suicide in opera has received only modest interest in the scientific literature. One qualitative study6 and one quantitative study7 are available, and these are both dated. The author of the qualitative study opined that, while suicide was common, “attempted suicides” were “missing from the operatic stage”.6 The quantitative study, published in 1971, was cross-sectional and found that, in leading opera houses throughout the world, suicide featured in 26% of the different operas and in 40% of the most popular operas.7

Our aim in this study was to examine the depiction of completed suicide as well as non-fatal suicidal acts and suicidal thought in Western opera over the past four centuries, with a view to characterising such behaviour as conceived by librettists and received by the public, and as an indication (at least in part) of more widespread contemporaneous beliefs about suicide.

We examined all of the operas that debuted from 1607 to 2006, inclusive, that were listed in a popular opera guide published in 2007.5 This guide provides a synopsis of each opera. If further details were required, we consulted comprehensive accounts.8–10 We considered the overall period 1607–2006 as well as separate 100-year periods (1607–1706, 1707–1806, 1807–1906 and 1907–2006). We collected data on completed suicides, non-fatal suicidal acts and suicidal thoughts, and on the circumstances of these and whether there was evidence of an underlying mental disorder. We recorded the sex of the characters involved and, for completed suicide, the method.

Although our main focus was suicidal behaviour and thought, we also recorded other forms of violence (such as war, murder, rape and kidnap), and identified operas in which murder preceded suicide.

We identified 337 operas in the period of interest. Over those 400 years, there was at least one suicide in 74 operas (22%), mostly among female characters, who accounted for 56% of cases. Non-fatal suicidal acts or suicidal thoughts were found in 48 operas (14%), predominantly among male characters cutting (including stabbing), who accounted for 57% of cases. In 112 operas (33%), there was completed suicide and/or non-fatal suicidal acts and suicidal thoughts. The suicide-related events in operas in the separate 100-year periods, and overall, are shown in Box 1.

Cutting (including stabbing) was the most common method of suicide in opera (26 cases); this was the case for both male (15) and female (11) characters. Other methods included poisoning (15 cases; six male, nine female), drowning (10 cases; two male, eight female), hanging (four cases; two male, two female), asphyxiation (four cases; all female), “supernatural” methods (four cases; two male, two female), immolation (three cases; two male, one female), jumping from a height (two cases; both female), shooting (one male), blunt trauma (one male) and refusing life-saving treatment (one female). Mass suicide occurred in two operas.

In 178 operas (53%), there were other forms of violence — murder, rape, torture and arson were not infrequent events — and it was noteworthy that many operas were set during periods of war. Murder preceded suicide in 12 operas.

Over the course of four centuries, a third of all of the operas we examined featured completed suicide, non-fatal suicidal acts or suicidal thoughts. Our finding that completed suicide was depicted in 22% of operas is consistent with rates of 25%6 and 26%7 reported in earlier studies. However, our finding that non-fatal suicidal acts or suicidal thoughts were represented in 14% of operas is in contrast to that of an earlier report that found no evidence of “attempted suicide” in opera.6 We did not separate non-fatal suicidal acts from suicidal thoughts, but these appear to have been represented about equally, and we identified at least a modest number of non-fatal suicidal acts by characters in operas (examples include Atys in Lully’s Atys, Idamante in Idomeneo, and Andronico in Handel’s Tamerlano).

It is intriguing that cutting was the most common method of completed suicide for both male and female characters, as this is not the case in real life (notwithstanding possible differences in method over time, and according to culture). The World Health Organization mortality database suggests that, from the 1950s, shooting, hanging, jumping from a height and poisoning have been the most frequent methods, and that violent methods are more frequent among men than women.11 Although we can not be certain why cutting is the most common method in opera, presumably it has greater dramatic potential onstage, which would appeal to both the librettist and the audience.

It is also of interest that female characters accounted for most completed suicides in opera (56%) and that non-fatal acts or thoughts were more common for male characters (57%); both findings are a reversal of the sex differences encountered in real life. Again, we can only speculate about the reasons, but it is arguable that this may have been influenced by the librettist’s and audience’s sex role perceptions. Completed suicide may have been somehow more socially acceptable for women than men, for whom it may have represented a sign of weakness.

Importantly, without exception, the suicide or non-fatal suicidal acts or suicidal thoughts described in our opera sample immediately followed an undesirable event or situation, such as loss of a loved person, loss of position or authority or imminent submission to a dominant force, rather than occurring in the context of an underlying mental disorder (as much as one can recognise such disorders from libretti). Familiar examples include Romeo’s suicide in Gounod’s Roméo et Juliette (Box 2) and Cio-Cio-San’s in Puccini’s Madama Butterfly (Box 3). It is possible that the range of symptoms and signs satisfying criteria for a mental disorder are often too difficult for the librettist to accurately and persuasively depict. Nevertheless, the notion that suicide does not invariably occur in the setting of mental disorder has been articulated in previous publications.12,13 Our own group has developed the concept of “predicament suicide”, in which we have delineated two major predicaments that may contribute to suicide.14,15 One predicament is untreated or unresponsive painful mental disorder and the other is an intolerable social situation. Completed suicide in opera appears to occur in response to predicaments in the latter category.

It is worth commenting briefly on the higher percentages of suicide and non-fatal suicidal acts and suicidal thoughts in the 18th and 19th centuries compared with the 20th and early 21st centuries. This finding could be attributed to changes in the attitude to suicide. In the early 18th century, suicide was considered a sin by the church and a crime under the law. Later that century, suicide shifted from a moral to a medical (mental illness) problem, and then came the “standard view” proposing that suicide was not always the result of mental disorder.11 The Romantic Movement favoured notions like suicide as an escape from a cruel world and a method of solving painful love problems. The new paradigm commenced about the time of publication of Goethe’s Die Leiden des jungen Werther (The sorrows of young Werther) in 177416 (which led to a large number of “copycat” suicides17) and the beginning of the French Revolution, and continued to the end of the 19th century. During the 20th century, suicide was strongly cast as a mental disorder rather than a reaction to a distressing event or situation.1,3,18–20 This may have reduced librettists’ use of suicide and related behaviour in operas over the past century.

Notwithstanding these historical fluctuations, what is clear is that over many centuries in Western and other cultures, suicide has been represented as an option when individuals have been faced with various forms of severe loss or other distressing situations. In our earlier work, we found numerous examples of suicide in response to stress, without apparent mental disorder, in the folk-stories of the ancient people of Greece21 and Germany and Finland,22 in satirical tales of 14th century Italy in Boccaccio’s The Decameron,23 and in more recent literary fiction.24

It would be fascinating to ascertain whether suicide rates in opera have varied according to country, in much the same way that suicide rates in the real world vary in different countries. However, this would be very difficult because librettists and composers are mobile people — many are born in one country, settle in another and have their work set in or first performed in yet another. Our study is essentially of the “opera of Europe”.

A limitation of our study is that only 337 operas were examined, and many thousands have been written. However, our sample was extracted, in toto, from a recent opera guide whose author, Matthew Boyden, set out to represent all eras of opera history, and “concentrated on the ones you’re likeliest to encounter either on disc or on stage”.5

Well meaning colleagues and musicians offered us lists of operas in which completed suicide, attempted suicide or suicidal thinking featured, but as these would have introduced bias we put them aside and included only those operas mentioned in Boyden’s text.

Apart from providing a glimpse into community beliefs about suicidal behaviours and data about changes in beliefs over time, our findings prompt speculation about the impact of opera depictions of suicide on real-world suicidal behaviour. There is ongoing debate about the influence of depictions of suicide in various media and the arts on subsequent copycat suicides.25 It may be too adventurous to suggest that the portrayal of suicide in opera will directly contribute to self-harming behaviour of members of the community on a large scale (let alone suggest that the drop in opera depictions of suicide in the 20th century was due to librettists’ potentially heightened social responsibility). Although not lacking exposure — it is estimated that there will be over 24 500 opera performances globally in the 2012–2013 season26 — opera is struggling to retain, let alone broaden its audience.27 Opera may not be as popular or potentially influential as, say, a blockbuster movie. Nevertheless, it is staggering to contemplate that a full third of operas written over a period of 400 years have featured completed suicide, non-fatal suicidal acts or suicidal thoughts. Might not these depictions affect the behaviour of a vulnerable member of the audience or a brittle person who can nowadays view opera in DVD and other formats? The “Papageno effect”, denoting preventive suicide action, is worth mentioning here. This term was derived from Mozart’s 1791 opera, Die Zauberflote (The magic flute), referring to a situation where Papageno, a young man in love, becomes suicidal on believing he has lost his love, but copes following his friends’ intervention (Box 4).28

In conclusion, the representation of suicide in opera is too prominent to be ignored. While many opera buffs may focus on the musical elements rather than the action and libretti of this art form, the depiction of suicide in operatic works adds to our understanding of the cultural dimensions of suicide over time, and thus to our overall understanding of this tragic outcome.

2 Deborah Riedel as Juliette and Stephen Ibbotson as Roméo in Gounod’s Roméo et Juliette

Libretto by Jules Barbier and Michel Carré.
1992 Opera Australia production. Photo: Kiren Chang.

3 Jennifer Barnes as Cio-Cio-San in Puccini’s Madama Butterfly

Libretto by Luigi Illica and Giuseppe Giacosa.
1999 Opera Australia production. Photo: Marco Bok.

4 Warwick Fyfe as Papageno in Mozart’s Die Zauberflote
(The magic flute)

Libretto by Emanuel Schikaneder.
2009 Opera Australia, production. Photo: Branco Gaico.